Archives of Clinical Neuropsychology | 2021

A-157 Lexical-Semantic Analysis of Verbal Naming Test Performance in Older Adult Veterans

 
 
 
 
 

Abstract


\n \n \n The Verbal Naming Test (VNT), a validated non-visual measure of word-finding ability in older adults, was designed to include words from distinct lexical (i.e., verbs vs. nouns) and semantic (i.e., action vs. object) categories to better capture various naming deficits. This study examined specific contributions of lexical and semantic categories to the convergent and discriminant validity of the VNT, and to how well it detects Mild Cognitive Impairment (MCI).\n \n \n \n The sample of older Veterans included 157 healthy controls (mean age\u2009=\u200974.23, SD\u2009=\u20096.67; range:60–89) and 39 diagnosed with MCI (mean age\u2009=\u200972.97, SD\u2009=\u20098.24; range:60–88). VNT item subscales were constructed for verbs/actions (ActionW) and nouns/objects presented either with (Obj+) or without (ObjW) an action component in the prompt. Correlations were calculated between VNT total/subscales and several other neuropsychological measures. ROC analyses compared relative contributions of each VNT subscale in detecting MCI.\n \n \n \n The ActionW subscale showed poor reliability (Cronbach’s alpha\u2009=\u20090.303) and did not correlate with any other measures. The ObjW, and Obj\u2009+\u2009subscales showed better reliability (Cronbach’s alpha>0.62) and correlated with age, measures of memory, and the Boston Naming Test. The Area Under the Curve (AUC) for detection of MCI was as follows: ActionW\u2009=\u20090.616; Obj\u2009+\u2009=0.635; ObjW\u2009=\u20090.689; and 50items\u2009=\u20090.690.\n \n \n \n Lexical and semantic categories differentially affect VNT detection of MCI in older veterans. Object words, especially those without associated action components, provide the best measurement scale. Although still clinically relevant in detecting MCI, the poor reliability of verbs/action words suggests revising the VNT may improve its psychometric properties. Future research should replicate these results in a clinical sample.\n

Volume None
Pages None
DOI 10.1093/arclin/acab062.175
Language English
Journal Archives of Clinical Neuropsychology

Full Text